zk-org/zk | ProgramBench

Plain text note-taking assistant

1,107

Generated Behavioral Tests

87.6%

Best Score

Claude Opus 5.5 (xhigh)

Results by Model

Hover for details · Click a model to open it · Drag to zoom the x-axis · Shift-drag to pan · The line marks the Pareto frontier (best score per cost)

30 runs

#	Model	Score	Cost	Calls
1	Claude Opus 5.5 (xhigh)	87.6%	$32.22	276	trace →
2	Claude Fable 5.1 (xhigh)	85.5%	$52.89	290	trace →
3	Muse Wonder (xhigh)	81.7%	$57.00	554	trace →
4	Muse Wonder Full	81.4%	TBA	218	trace →
5	GPT 6 Astra (max)	75.0%	$9.30	60	trace →
6	Claude Opus 5 (xhigh)	64.1%	$24.42	187	trace →
7	Muse Spark 1.3 (xhigh)	61.3%	$1.86	285	trace →
8	Claude Opus 4.8 (xhigh)	59.7%	$20.06	163	trace →
9	Muse Spark 1.3 (max)	57.6%	$5.08	292	trace →
10	GLM-5.2	51.6%	$31.48	271	trace →
11	Gemini 3.7 Flash	43.6%	$1.79	133	trace →
12	Claude Sonnet 4.6	42.9%	$24.25	494	trace →
13	Muse Spark 1.1 (xhigh)	40.0%	$0.53	157	trace →
14	Gemini 3.6 Flash	38.8%	$4.31	191	trace →
15	GPT-5.6 Sol	26.0%	$0.66	12	trace →
16	GPT-5.6 Sol (xhigh)	22.9%	$5.09	34	trace →
17	Gemini 3.5 Flash	18.0%	$6.31	234	trace →
18	GPT 5.4	17.0%	$0.30	8	trace →
19	GPT 5.5 (high)	16.5%	$4.05	51	trace →
20	GPT 5.5 (xhigh)	16.3%	$7.05	80	trace →
21	Claude Opus 4.7 (xhigh)	16.1%	$9.09	113	trace →
22	GPT 5.5	15.3%	$1.30	22	trace →
23	Claude Opus 4.7	14.0%	$1.52	59	trace →
24	Gemini 3 Flash	13.9%	$0.21	62	trace →
25	Gemini 3.1 Pro	12.1%	$1.75	139	trace →
26	GPT 5 mini	7.7%	$0.02	12	trace →
27	GPT 5.4 mini	5.1%	$0.03	9	trace →
28	Muse Spark 1.2 (xhigh)	1.2%	$1.75	229	trace →
29	Claude Opus 4.6	1.2%	$14.07	286	trace →
30	Claude Haiku 4.5	1.2%	$0.59	91	trace →

Click a row to replay how that model rebuilt this program, or the model name to open its full run